[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Re : E-journals in the era of print cancellations

I wrote:

>> We deliver both HTML (generated from the native SGML database) and
>> PDF, and while the lower resolution GIF figures in the HTML version
>> look OK on the screen and the PDF versions print out about as well
>> as a good photocopy, neither rivals the higher resolution images
>> present in the print journal.

and Mark Steinberger replies:

>So why not include links to higher resolution versions in the html
>or pdf files? You could have a link directly from the lower resolution
>in-line version.

This was discussed at great length this summer by the group of scientists
that advises us on matters related to effective use of electronic
publications within our community (and, hence, helps keep us on track as
we move forward).

The issue on the surface is this: we have a great number of images in our
journals and to store all of them in a resolution comparable to that used
for print would require massive storage.  In reality, the community would
not need ALL of them in high resolution but it would be nearly impossible
to decide a priori which ones they would need and which they would not
(especially when you think in the 10, 50, or 100 year time frame as you
must if you are saying this repository serves as the true archive for the
journal), so you would have to store all -- at significant cost and great
inefficiency of resources.

The issue at the heart of the matter is more subtle.  The sort of imagery
we deal with comes from many sources and is often the result of multiple
layers of processing.  Almost all of it is still provided by the authors in
the form of hardcopy that is scanned for reproduction in print and online.
In order to faithfully reproduce in print the original imagery, it must be
scanned at very high resolution.  In most cases, the "native" information
in the image has an intrinsic resolution that is much lower but the scan
resolution must be sufficiently high to ensure faithful capture of that
"native" information -- resulting in an electronic file that is many times
larger than would have been implied by the resolution of the "native"
information it contains.

The first step as we see it, is to work toward having as much as possible
of the imagery we reproduce come to us from authors as electronic files
that are at the resolution required by the native information they
contain. This reduces the problem but does not come close to eliminating
it, because it really is still more fundamental that this.  In many cases,
visualization software will create a rendering as a postscript image that
is a few megabytes in size from a dataset that is only tens of kilobytes
-- again there is a great inefficiency in storing the visualization of the
data instead of the data itself.  The technology is not sufficiently
standardized for us to be able to deliver just the data and let the
visualization take place on the client side.

Thus, our approach has been the following:

1) For now, continue to treat the print version as the "archive" version of
the journal to ensure that the image quality is preserved in the archive.
[Knowing, however, that this sets on hold our plans to declare the
electronic version as the "true" journal and incorporate "nonprintable"
material (animations, etc.) as integral components of journal articles.  We
currently deliver this sort of content as "supplemental material" on
CD-ROMs delivered with the print and also available online.]

2) Work toward a plan for retaining high resolution electronic images of
the figures that can be linked to from the inline images in the online
delivery, but not so aggressively that we create a massive database of
inefficiently stored material. The real plan here is to come up with the
procedures now for this approach so that it can be implemented as storage
and bandwidth availability make it more practical in a couple of years.

3) Anticipate a future when more of the visualization of the data can occur
on the client side so that the journal archive can store the true "native"
information and the usefulness of the data is enhanced by users being able
to manipulate it in addition to visualize it.

I should point out that all I have been discussing relates to our
traditional, longstanding, print journals that are now also delivered
online.  We are a copublisher of an all-electronic journal (Earth
Interactions) that already fully integrates data and nonprintable content
as we expect our traditional journals to do in the future. (And from this,
we are learning a lot.)

This has been, perhaps, a longer and more technical answer to Mark's
comment than is warranted on a listerv devoted to licensing issues, but as
it relates to the reasons for not replacing print with online at this time
in the evolution of the journal, I hope you all will forgive me.


Dr. Keith L. Seitter               phone:  617-227-2426 ext. 220
Associate Executive Director       fax:    617-742-8718
American Meteorological Society    e-mail: kseitter@ametsoc.org
45 Beacon Street
Boston, MA 02108-3693              http://www.ametsoc.org/AMS